Smol+async global executor 1.80 dev #3791

martin-kolarik · 2025-03-17T22:11:11Z

async-global-executor added as the fourth executor. The PR extends PR #3790.

This is a breaking change -- async-global-executor has MSRV 1.80, therefore sqlx will have to be upgraded too.

# Conflicts: # .github/workflows/sqlx.yml

abonander · 2025-03-17T22:14:01Z

I would prefer to deprecate async-std for a release cycle before deleting it. Just cause it's discontinued doesn't mean people will stop using it immediately, especially because its constituent crates will continue to receive updates.

martin-kolarik · 2025-03-17T22:15:21Z

I would prefer to deprecate async-std for a release cycle before deleting it. Just cause it's discontinued doesn't mean people will stop using it immediately, especially because its constituent crates will continue to receive updates.

Sure, I did not remove async_std from the sources, for the same reason.

abonander

Same note as here: #3790 (review)

# Conflicts: # Cargo.toml

abonander · 2025-07-05T00:46:03Z

@martin-kolarik do you plan on addressing the CI failure and merge conflicts?

martin-kolarik · 2025-07-05T21:31:03Z

@martin-kolarik do you plan on addressing the CI failure and merge conflicts?

I would like to address both. I will look at merge conflict first (today or on Monday) and we will see how tests changed. CI failure is linked to changes around to_socket_addrs and I suspect tests fail due to changed behavior. I will write more later.

# Conflicts: # Cargo.lock # Cargo.toml # rust-toolchain.toml # sqlx-cli/Cargo.toml # sqlx-core/Cargo.toml # sqlx-core/src/rt/mod.rs # sqlx-macros-core/src/lib.rs # sqlx-sqlite/src/connection/worker.rs

… selected

martin-kolarik · 2025-07-06T20:32:10Z

So, merged, fixed, and tests are running: the main issue of previous failures disappeared by itself, there certificate for localhost was expected.

CI completion is prevented with error in build using minimal versions check, where async-fs fails. According to rust-lang/cargo#5657 it may happen. Did you have similar issues in the past? How to fix it?

abonander · 2025-07-07T01:32:53Z

Strictly speaking, this is a bug in async-fs 2.0.0 because they're calling futures_lite::FutureExt::poll() which did not exist yet in the version they actually specify (1.2).

The whole idea of the minimal-versions test is to make sure that SQLx still builds if someone hasn't updated some arbitrary dependency below it in the tree, so the test is working as expected here. However, this isn't strictly our problem to begin with.

We could switch to -Z minimal-direct-versions which only lowers our direct dependencies to their minimum specified versions, so we'd only be testing our own specifications. This gives limits the scope to dependencies we actually control, but also weakens the guarantees of the minimal-versions check (since indirect dependencies could still be technically broken).

I'd really like feedback from @iamjpotts here since he's the one who recommended adding the check to begin with (#3340).

iamjpotts · 2025-07-07T03:20:58Z

The whole idea of the minimal-versions test is to make sure that SQLx still builds if someone hasn't updated some arbitrary dependency below it in the tree, so the test is working as expected here. However, this isn't strictly our problem to begin with.

Yes, it is intended to be a correctness check on the declared minimums.

From that standpoint the ci check is working as designed. The original issue was mentioned in #3118 (linked from #3340).

async-fs fixed their dependency declaration in async-fs 2.1.

smol requires futures-lite 2.0, but in sqlx is optional.

The fix is likely to be adding an explicit dependency on a transitive dependency, where the explicit dependency enforces the correct minimum. Possible workarounds to explore:

Add a sqlx dependency on async-fs 2.1
Add a sqlx dependency on futures-lite 2.0

For either it would be a good idea to add a comment that the above is to resolve a problem with a transitive dependency (and which one, idk if it is smol without further investigation)

Try to fix CI -Z minimal-versions check

martin-kolarik · 2025-07-07T18:26:35Z

Add a sqlx dependency on async-fs 2.1

Add a sqlx dependency on futures-lite 2.0

I added dependency on async-fs 2.1, as it imo better describes the intention of the workaround: the async-fs 2.0 contains bug, not futures-lite.

iamjpotts · 2025-07-07T18:50:15Z

Add a sqlx dependency on async-fs 2.1

Add a sqlx dependency on futures-lite 2.0

I added dependency on async-fs 2.1, as it imo better describes the intention of the workaround: the async-fs 2.0 contains bug, not futures-lite.

This specific change lgtm though @abonander may have more fine grained feedback.

iamjpotts · 2025-08-10T18:41:36Z

sqlx-core/src/rt/mod.rs

    }

-    panic!("either the `runtime-async-std` or `runtime-tokio` feature must be enabled")
+    panic!("one of the `runtime-async-global-executor`, `runtime-async-std`, `runtime-smol`, or `runtime-tokio` feature must be enabled")


Grammar - should be features

Sure, should I change PR?

iamjpotts · 2025-08-10T18:42:37Z

sqlx-core/src/net/socket/mod.rs

+// Work around `impl Socket`` and 'unability to specify test build cargo feature'.
+// `connect_tcp_address` compilation would fail without this impl with
+// 'cannot infer return type' error.
+impl Socket for () {


Were any alternatives to this considered?

Why is this unreachable?

impl Socket for () satisfies Rust type system, it is necessary at @260, when neither _rt-tokio nor _rt-async-io is selected. rustc cannot resolve impl Socket type without defining any. Also that's why the impl uses unreachable!, it is erased and unused during runtime.

The code has been added after discussing it in #3790 (comment), where shared implementation for both executors to resolve host names was suggested. I think it is fine, but it ended with impl Socket issue.

IMO there is alternative to use enum on connect_tcp_address return, but I expect it would bubble up through connect_tcp, which also has generic output, so enum would appear in many places. Using void impl on unit type seemed better to me.

abonander · 2025-09-06T01:20:36Z

sqlx-core/src/net/socket/mod.rs

+
+                Ok(stream)
+            } else {
+                crate::rt::missing_rt(socket_addr)


This is actually a breaking behavior change. The existing behavior is to fall back to async_io if a Tokio runtime isn't available, which could be important for some use-cases.

Because of Cargo's feature unification, if a library using SQLx turns on the runtime-tokio feature, it's forced on for users of that library as well, whether they want it or not. The only way to turn it off is to patch the library.

This change would make a situation like that a hard blocker instead of mildly annoying.

abonander · 2025-09-08T18:17:50Z

I'm going to merge to avoid the need for any more back-and-forth, but I'm going to be making some adjustments in a follow-up PR.

martin-kolarik · 2025-09-08T18:34:55Z

I'm going to merge to avoid the need for any more back-and-forth, but I'm going to be making some adjustments in a follow-up PR.

Thank you, unfortunately I was off during weekend so I did not respond to breaking change note. Feel free to ask for anything to be completed.

* restore fallback to `async-io` for `connect_tcp()` when `runtime-tokio` feature is enabled * `smol` and `async-global-executor` both use `async-task`, so `JoinHandle` impls can be consolidated * no need for duplicate `yield_now()` impls * delete `impl Socket for ()`

* Sqlx-core: rename async_io dependency for async-std * Sqlx-core: simplify TimeoutError * Sqlx-core: code for async-global-executor * Sqlx: integrate async-global-executor feature * Note to unsafe * Step up MSRV as async-global-executor needs it * Sqlx-core: fix of unix socket build * Unsafe fixes, smol executor added * Workflow fix * Changes outside sqlx_rt * Cleanup conditional rt compilation * Warning * Add executors to test matrix * Fix of skipping code sqlite due to mismatch in cargo feature names * Smol executor isolated * Fix, reduce number of tests, remove async_std * Fix of test_block_on, regression * Format fixes * async-global-executor added * async-std changed to 1.13 * litemap, zerofrom requires rustc 1.81 * Fix of missing _sqlite in cargo.toml * Clippy lints * Clean up * Remove features combinations * Fixes after merge * Fix of compiling connect_tcp_address with both tokio + other executor selected * Try to fix CI -Z minimal-versions check Try to fix CI -Z minimal-versions check

* restore fallback to `async-io` for `connect_tcp()` when `runtime-tokio` feature is enabled * `smol` and `async-global-executor` both use `async-task`, so `JoinHandle` impls can be consolidated * no need for duplicate `yield_now()` impls * delete `impl Socket for ()`

martin-kolarik added 22 commits March 2, 2025 15:11

Sqlx-core: rename async_io dependency for async-std

fb3f254

Sqlx-core: simplify TimeoutError

31553e7

Sqlx-core: code for async-global-executor

0c12744

Sqlx: integrate async-global-executor feature

6c324b1

Merge with upstream

46f8400

Note to unsafe

40f3d8d

Step up MSRV as async-global-executor needs it

375b493

Sqlx-core: fix of unix socket build

aef441b

Unsafe fixes, smol executor added

f5e374a

Workflow fix

7de5718

Changes outside sqlx_rt

b462f43

Cleanup conditional rt compilation

4f851b0

Warning

4eadb24

Merge branch 'temp'

dd053e1

# Conflicts: # .github/workflows/sqlx.yml

Add executors to test matrix

aa29a3e

Fix of skipping code sqlite due to mismatch in cargo feature names

89faf0e

Merge remote-tracking branch 'origin/sqlx-cli-sqlite-create-WAL-db-fix'

8cc0f5c

Smol executor isolated

ddcf4ed

Fix, reduce number of tests, remove async_std

653152a

Fix of test_block_on, regression

d1e91f6

Format fixes

5627ca8

async-global-executor added

c7fa793

abonander added the breaking:MSRV Requires waiting for an MSRV bump (see FAQ.md) label Apr 11, 2025

abonander reviewed Apr 12, 2025

View reviewed changes

martin-kolarik and others added 3 commits April 13, 2025 09:49

Merge branch 'launchbadge:main' into main

580b8df

Merge branch 'main' into smol+async-global-executor-1.80-dev

1698c3b

async-std changed to 1.13

8d6377b

martin-kolarik force-pushed the smol+async-global-executor-1.80-dev branch from 1f43f07 to 8d6377b Compare April 13, 2025 11:39

Clean up

9b0cebb

martin-kolarik force-pushed the smol+async-global-executor-1.80-dev branch from 8c2c83c to 9b0cebb Compare April 13, 2025 13:50

Remove features combinations

c3e8f61

martin-kolarik force-pushed the smol+async-global-executor-1.80-dev branch from 35237ac to c3e8f61 Compare April 13, 2025 14:24

Merge branch 'main' into smol+async-global-executor-1.80-dev

2d41986

# Conflicts: # Cargo.toml

martin-kolarik added 3 commits July 6, 2025 19:02

Merge branch 'main' into smol+async-global-executor-1.80-dev

8afee18

# Conflicts: # Cargo.lock # Cargo.toml # rust-toolchain.toml # sqlx-cli/Cargo.toml # sqlx-core/Cargo.toml # sqlx-core/src/rt/mod.rs # sqlx-macros-core/src/lib.rs # sqlx-sqlite/src/connection/worker.rs

Fixes after merge

1fe4472

Fix of compiling connect_tcp_address with both tokio + other executor…

2d1d910

… selected

Try to fix CI -Z minimal-versions check

3b95c99

Try to fix CI -Z minimal-versions check

martin-kolarik and others added 2 commits July 8, 2025 13:36

Merge branch 'main' into smol+async-global-executor-1.80-dev

c638898

Merge branch 'main' into smol+async-global-executor-1.80-dev

a641111

iamjpotts reviewed Aug 10, 2025

View reviewed changes

abonander mentioned this pull request Aug 27, 2025

PostgreSQL SASL – run SHA256 in a blocking executor #4006

Merged

Merge branch 'main' into smol+async-global-executor-1.80-dev

db2cecf

abonander reviewed Sep 6, 2025

View reviewed changes

abonander merged commit 6b828e6 into launchbadge:main Sep 8, 2025
144 checks passed

Uh oh!

Smol+async global executor 1.80 dev #3791

Smol+async global executor 1.80 dev #3791

Uh oh!

Conversation

martin-kolarik commented Mar 17, 2025

Uh oh!

abonander commented Mar 17, 2025

Uh oh!

martin-kolarik commented Mar 17, 2025

Uh oh!

abonander left a comment

Choose a reason for hiding this comment

Uh oh!

abonander commented Jul 5, 2025

Uh oh!

martin-kolarik commented Jul 5, 2025

Uh oh!

martin-kolarik commented Jul 6, 2025

Uh oh!

abonander commented Jul 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

iamjpotts commented Jul 7, 2025

Uh oh!

martin-kolarik commented Jul 7, 2025

Uh oh!

iamjpotts commented Jul 7, 2025

Uh oh!

iamjpotts Aug 10, 2025

Choose a reason for hiding this comment

Uh oh!

martin-kolarik Aug 10, 2025

Choose a reason for hiding this comment

Uh oh!

iamjpotts Aug 10, 2025

Choose a reason for hiding this comment

Uh oh!

martin-kolarik Aug 11, 2025

Choose a reason for hiding this comment

Uh oh!

abonander Sep 6, 2025

Choose a reason for hiding this comment

Uh oh!

abonander commented Sep 8, 2025

Uh oh!

Uh oh!

martin-kolarik commented Sep 8, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

abonander commented Jul 7, 2025 •

edited

Loading